A metal-free photocatalyst for highly efficient hydrogen peroxide photoproduction in real seawater

Page created by Edith Baker
 
CONTINUE READING
A metal-free photocatalyst for highly efficient hydrogen peroxide photoproduction in real seawater
ARTICLE
                   https://doi.org/10.1038/s41467-020-20823-8               OPEN

                  A metal-free photocatalyst for highly efficient
                  hydrogen peroxide photoproduction in real
                  seawater
                  Qingyao Wu1, Jingjing Cao1, Xiao Wang1, Yan Liu1, Yajie Zhao1, Hui Wang1, Yang Liu1 ✉, Hui Huang1, Fan Liao1,
                  Mingwang Shao1 ✉ & Zhenghui Kang 1,2 ✉
1234567890():,;

                  Artificial photosynthesis of H2O2 from H2O and O2, as a spotless method, has aroused
                  widespread interest. Up to date, most photocatalysts still suffer from serious salt-deactivated
                  effects with huge consumption of photogenerated charges, which severely limit their wide
                  application. Herein, by using a phenolic condensation approach, carbon dots, organic dye
                  molecule procyanidins and 4-methoxybenzaldehyde are composed into a metal-free photo-
                  catalyst for the photosynthetic production of H2O2 in seawater. This catalyst exhibits high
                  photocatalytic ability to produce H2O2 with the yield of 1776 μmol g−1h−1 (λ ≥ 420 nm;
                  34.8 mW cm−2) in real seawater, about 4.8 times higher than the pure polymer. Combining
                  with in-situ photoelectrochemical and transient photovoltage analysis, the active site and the
                  catalytic mechanism of this composite catalyst in seawater are also clearly clarified. This
                  work opens up an avenue for a highly efficient and practical, available catalyst for H2O2
                  photoproduction in real seawater.

                  1 Institute
                            of Functional Nano and Soft Materials Laboratory (FUNSOM), Jiangsu Key Laboratory for Carbon-Based Functional Materials & Devices, Soochow
                  University, 215123 Suzhou, PR China. 2 Macao Institute of Materials Science and Engineering, Macau University of Science and Technology, Taipa, 999078
                  Macau SAR, China. ✉email: yangl@suda.edu.cn; mwshao@suda.edu.cn; zhkang@suda.edu.cn

                  NATURE COMMUNICATIONS | (2021)12:483 | https://doi.org/10.1038/s41467-020-20823-8 | www.nature.com/naturecommunications                             1
A metal-free photocatalyst for highly efficient hydrogen peroxide photoproduction in real seawater
ARTICLE                                                                 NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-20823-8

H
         ydrogen peroxide (H2O2) has aroused widespread interest              Results
         as a high-energy oxidant for the potentially replacement             Characterizations of CDs. Firstly, the structure and photoelec-
         of fossil fuel because of its widely application in the field         trochemical properties of CDs were studied. Figure 2a shows
of medicine, chemical industry, and environmental manage-                     the transmission electron microscopy (TEM) image and particle
ment1–3. The traditional approaches for H2O2 production, such                 size distribution taken from monodispersed CDs, making clear
as anthraquinone method and electrochemical synthesis, result in              that CDs are spherical nanoparticles with a diameter of about
huge energy consumption and high toxicity by-products4,5.                     6–16 nm. A high-resolution TEM (HRTEM) image displayed in
Artificial photosynthesis method, namely the reaction of oxygen                Fig. 2b exhibits a lattice spacing of 0.21 nm, which corresponds to
and water driven by sunlight, is recognized as the most promis-               (100) lattice planes of graphitic carbon19. The functional groups
ing, spotless method for the H2O2 green fabrication6–8. Given the             of CDs are indicated in the Fourier transform infrared (FT-IR)
abundance of seawater, the utilization of seawater to photo-                  spectrum (inset in Fig. 2b), where the peaks located at 1632, 1397,
produce H2O2 can not only mitigate the problem of fresh water                 and 1233 cm−1 are assigned to the stretching vibration of C=O,
resource shortage but also greatly reduce the cost of reaction                –COO, and C–OH25, respectively. Power X-ray diffraction (XRD)
system process9,10. Although more and more materials are being                pattern of CDs in Fig. S1 shows two broad peaks at around 21°
used in photocatalytic synthesis of H2O2, most of them are still              and 43° corresponding to the (002) and (100) planes of graphitic
limited by such factors as poor light absorption11,12, mismatched             carbon.
band gap13,14, low electron transfer rate15, and essential sacrificial            The full X-ray photoelectron spectroscopy (XPS) spectrum in
agents16,17. In particular, salt can deactivate the catalysts or              Fig. S2a demonstrates only C and O elements in CDs. The high-
consume photogenerated carriers, leading to undesirable side                  resolution spectrum of C 1s can be fitted for C–C/C=C, C–O, and
reactions on the surface of the catalyst, which severely limits the           C=O while the O 1s can be matched to Osurf and Oads26,27.
industrialization process of H2O2 production by sunlight18,19.                Further structure analysis on CDs is shown in Fig. 2c, where the
   In photocatalytic field, carbon dots (CDs) have been proved to              hydrion in the carboxyl group shows electropositivity (δ1+) while
be the co-catalytic active site and/or the good electron acceptor/            the adjacent oxygen is electronegativity (δ1−). When the metal
donor material, and show unique ability to improve the catalytic              salt (M+) is added, the carboxyl group can be ionized, and the
efficiency in the photocatalytic system20–23. Herein, as shown in              electronegativity (δ2–) of oxygen in the carbon group increase,
Fig. 1, we report a phenolic condensation approach, in which CDs,             which will, in turn, increase the charges and electron sink barrier
organic dye molecule procyanidins, and 4-methoxybenzaldehyde                  of CDs, and further prolongs the life of the electron. Density
are composed into metal-free photocatalyst (PM-CDs-x) for the                 functional theory calculations have been conducted to understand
photocatalytic production of H2O2 in real seawater. Notably, we               the effect of ions on the electrostatic potentials of CDs with
propose and prove the electron sink effect of CDs, and this elec-             respect to the energy level of a vacuum. As shown in inset images
tron sink effect would grow with the addition of metal cations. So,           of Fig. S3, CDs are modeled with one-dimensional graphene
the catalyst can extract more electrons under light excitation,               nanoribbons, the edge of which are functionalized with carboxyl
effectively hindering the electron–hole recombination. It turns out           (−COOH) and sodium carboxylate (−COONa). The computa-
that this ideal composite catalyst exhibits the expected photo-               tional details can be found in the supporting information. The
catalytic capacity to generate H2O2 in seawater, and the catalytic            work function of these graphene nanoribbons can be estimated by
activity of the optimal composite photocatalyst is approximately              plane-averaged potentials. Figure S3 shows the plane-averaged
4.8 times high than the pure polymer photocatalyst in seawater. In            potentials of nanoribbons are plotted along the direction that
particular, the maximum yield of H2O2 for the optimal catalyst                perpendicular to the surface. It is seen that the trapping of
PM-CDs-30 is 1776 μmol g−1 h−1 in seawater, the apparent                      electrons is more profound in the graphene nanoribbon
quantum yield (AQY) is as high as 0.54% at 630 nm in seawater,                terminated with carboxyl groups. This is in good agreement with
and the solar-to-chemical conversion (SCC) efficiency can reach                our predictions.
0.21% in seawater. Combining with H2O2 generation rates, in situ                 The TPV experiments were carried out and shown in Fig. 2d, in
photoelectrochemical and transient photovoltage (TPV) tests, the              which the TPV curve of CDs decays slowly and gently, indicating a
role of CDs in the composite catalyst PM-CDs-30, during                       slow combination of charges. Obviously, after the addition of NaCl,
the photocatalytic reaction process, was established, and the active          the intensity of photovoltage increased, which is mainly because
site and the working mechanism of the composite catalyst in                   that the enhanced electron-withdrawing property of the functional
seawater were also clarified. Moreover, as suggested in a recent               group by ionization increases electron extraction rate under the
article24, the optimal reaction conditions of the catalyst were               same excitation. Furthermore, due to the increased energy barrier
investigated through experiments and thermodynamic–kinetic                    of the electron sink, the “trapped” electrons surrounding adjacent
model. Our results open up an avenue for seawater utilization,                oxygen cannot participate in the electron–hole recombination
catalyst design, and H2O2 generation.                                         process in a short time. Therefore, CDs draw electrons from the

Fig. 1 Synthesis process. The synthesis process of organic polymer composite systems. PM dipolymer represents procyanidins-methoxybenzaldehyde
diploymer.

2                NATURE COMMUNICATIONS | (2021)12:483 | https://doi.org/10.1038/s41467-020-20823-8 | www.nature.com/naturecommunications
A metal-free photocatalyst for highly efficient hydrogen peroxide photoproduction in real seawater
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-20823-8                                                                      ARTICLE

Fig. 2 Structure and property of CDs. a TEM image of CDs with the particle size distribution inserted. b HRTEM image of CDs with FT-IR spectrum
inserted. c Schematic diagram of the electron sink model for CDs. The black boll represents a single CD, M+ represents metal cations. d TPV curves of CDs
powders before and after adding NaCl.

system, which causes negative signal region on the TPV curve (the               enhancement effect, suggesting that it can be used as the ORR site
red line in Fig. 2d). As time goes on, the “trapped” electrons slowly           in the catalytic system. Thus, the functional groups (such as –OH,
pass through the barrier and recombine with holes, eventually                   C=O, –COOH) on the surface of CDs not merely play an
reaching equilibrium. The TPV tests of CDs mixed with other                     extraordinary role in the design and performance regulation of
metal salt ions shown in Fig. S5 display the same phenomenon as                 catalysts, but also effectively trap electrons through its electron sink
that added with NaCl. Another important property in Fig. S6                     effect in seawater to accelerate the electron transfer and prevent
reveals that CDs can perform electrocatalytic oxygen reduction                  electron–hole recombination, thereby improving the photocatalytic
reaction (ORR) by two-electron channel with no photoelectric                    activity of the catalyst in seawater.

NATURE COMMUNICATIONS | (2021)12:483 | https://doi.org/10.1038/s41467-020-20823-8 | www.nature.com/naturecommunications                                3
A metal-free photocatalyst for highly efficient hydrogen peroxide photoproduction in real seawater
ARTICLE                                                                 NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-20823-8

Fig. 3 Structure and property of PM-CDs-30. a TEM image and HRTEM image (inset), b UV–vis absorption spectra (inset: the corresponding Tauc plot),
c UPS spectrum, and d band structure diagram of PM-CDs-30.

Characterizations of composite photocatalysts. PM-CDs-x                       The switched Tauc plot of (αhν)1/2 versus photon energy (hν)
was prepared by adding CDs into the crosslinked polymerization                is obtained according to the formula introduced in the previous
of procyanidins-methoxybenzaldehyde (PM) dipolymer. Here,                     works12, and the band gap energy (Eg) of PM-CDs-30 is
PM-CDs-30 is taken as an example to show the structure of                     determined to be 1.94 eV. Then, UPS is used to estimate the
the photocatalyst. Scanning electron microscope (SEM) image of                ionization potential (equal to the valence band energy (EVB)) and
PM-CDs-30 (Fig. S7a) and its partial enlarged view (Fig. S7b)                 investigate the band position of PM-CDs-30. According to Fig. 3c,
indicate homogeneous micron-sized particles with flake accu-                   by subtracting the width of the He I UPS spectrum from the
mulation, while the SEM image of PM-CDs-0 (Fig. S7c) shows                    excitation energy (21.22 eV), the EVB of PM-CDs-30 is deter-
nonuniform spherical particles. Further observation of the                    mined to be 5.98 eV (vs. vacuum). Thus, the conduction band
structure through TEM (Fig. 3a) shows that CDs with the size of               (ECB) is calculated to be 4.04 eV (vs. vacuum). Visualizing, the
6–16 nm are uniformly distributed in the lamellar layer of the                energy level diagram of PM-CDs-30 is displayed in Fig. 3d,
polymer PM-CDs-30. A HRTEM image inserted in Fig. 3a gives a                  indicating that the energy level of the conduction band of PM-
lattice spacing of 0.21 nm for the single CD in PM-CDs-30,                    CDs-30 is higher than that of the ORR (0.68 V vs. RHE). Noting
corresponding to the (100) lattice planes of graphitic carbon. In             that the energy level of the valence band is located below the
addition, a series of basic characterizations, such as the size dis-          oxidation level for O2/H2O (1.23 eV vs. RHE) but above than that
tribution (Fig. S8), XRD patterns (Fig. S9a), FT-IR spectra                   for H2O2/H2O (1.78 eV vs. RHE). Thence, the organic polymer
(Fig. S9b), XPS spectra (Fig. S10), elemental analysis (Table S1),            composite PM-CDs-30 with appropriate band position can realize
and specific surface area (Table S2), declare that PM-CDs-30 is an             the ORR and water oxidation reaction simultaneously.
amorphous carbon structure composed of C, H, O elements with
abundant functional groups.
   The optical properties and band structure of PM-CDs-30 were                Photocatalytic activity of the composite systems PM-CDs-x.
examined by ultraviolet–visible (UV–vis) absorption spectro-                  The photocatalytic activities of the as-prepared photocatalysts
scopy and ultraviolet photoelectron spectroscopy (UPS) test.                  were studied in pure water and real seawater at room tempera-
UV–vis adsorption spectra in Fig. 3b demonstrate that PM-CDs-                 ture and pressure, and the results are summarized in Fig. 4a.
30 shows well absorption performance in ultraviolet, visible, and             Obviously, the pure polymer PM-CDs-0 generates H2O2 with a
even near-infrared region. It is worth noting that the addition of            low production rate of 771 μmol g−1 h−1 in pure water. While,
CDs enhances the absorption of photocatalyst at the wavelength                when the catalytic environment is changed to seawater, the
greater than 670 nm. As the content of CDs increases, the color of            activity of the catalyst reduced greatly, which is about 0.48 times
the catalyst deepened (Fig. S11), and the absorption in the near-             that of pure water. This is mainly due to the large number
infrared region also gradually increases, which is more conducive             of ionic components and impurities in seawater seriously
to the utilization of sunlight (Fig. S12). The optical band gap               affected the intrinsic structure and electron transport of
of PM-CDs-30 can be calculated from the Tauc plot (inset                      photocatalysts19,28,29. Relatively speaking, PM-CDs-30 shows the
in Fig. 3b) derived from the UV–vis absorption spectra.                       best activity in both pure water and seawater solution among the

4                NATURE COMMUNICATIONS | (2021)12:483 | https://doi.org/10.1038/s41467-020-20823-8 | www.nature.com/naturecommunications
A metal-free photocatalyst for highly efficient hydrogen peroxide photoproduction in real seawater
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-20823-8                                                                   ARTICLE

Fig. 4 Catalytic performance of PM-CDs-30. a Comparison of H2O2 production among the photocatalysts with different CDs contents in pure water and
real seawater. b Comparison of hydrogen peroxide yield rate between seawater and water with different catalysts. c The hydrogen peroxide production
ratio with different CDs contents to pure polymer catalyst PM-CDs-0. d Calculation and experimental on the dependence of H2O2 production rate of PM-
CDs-30 and rotational speed in different medium (red line: calculation data; black point: experimental data). e Calculation and experimental on the
dependence of H2O2 production rate of PM-CDs-30 and oxygen partial pressure in different medium (red line: calculation data; black point: experimental
data). f Time course of H2O2 photoproduction by PM-CDs-30. g Stability of photoproduction of H2O2 by PM-CDs-30. h Wavelength-dependent AQY of
oxygen reduction reaction by PM-CDs-30. The vertical error bars indicate the maximum and minimum values obtained; the dot represents the
average value.

catalysts studied in this work, and its photocatalytic activity in              for photocatalytic ORR, especially play an important role in the
seawater (1776 μmol g−1 h−1) is much higher than that in pure                   photocatalytic reaction under seawater condition. In terms of
water (1340 μmol g−1 h−1). And the SCC efficiency of PM-CDs-                     ORR photocatalytic activity, PM-CDs-30 composite was deter-
30 can reach 0.21% in seawater. To further explore the photo-                   mined to be an optimal ratio of CDs in this composite.
catalytic activity of the catalysts in real seawater and pure water,               The thermodynamic–kinetic model was performed to predict
the difference in hydrogen peroxide production rate between                     the optimum reaction conditions and maximum reaction rate of
seawater and pure water is displayed in Fig. 4b. Surprisingly, the              PM-CDs-30. Considering various factors that affect reactions
polymers with CDs added all show improved photocatalytic                        rates during the photocatalytic reactions, the rate optimization is
activity in seawater compared with pure water, and its photo-                   done with freely varying lattice parameters by using electro-
catalytic activity in seawater is about 1.1–1.5 times than that of in           chemical methods and in situ TPV tests. The detailed calculation
pure water. Importantly, Fig. 4c clearly shows the effect of the                process is shown in the supporting information. Here, we
added CDs on the photocatalytic performance of the composite                    consider the influence of water oxidation, oxygen reduction, and
catalyst, indicating that the addition of CDs can increase the                  oxygen diffusion on the reaction rate, and obtained the relation-
catalytic activity about by up to 4.8 times compared with pure                  ship between oxygen partial pressure, rotational speed, and
polymer PM-CDs-0 in real seawater. In summary, CDs are useful                   reaction rate. First, for oxygen evolution reaction (OER) in

NATURE COMMUNICATIONS | (2021)12:483 | https://doi.org/10.1038/s41467-020-20823-8 | www.nature.com/naturecommunications                             5
A metal-free photocatalyst for highly efficient hydrogen peroxide photoproduction in real seawater
ARTICLE                                                                          NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-20823-8

seawater, the reaction rate r1 can be obtained.                                        maximum rate is far lower than that in seawater, which is
                  þ r1
                                                                                       consistent with the experimental results in Fig. 4a. As shown in
    2H2 O þ 4h ! O2 þ 4Hþ                                                              Fig. 4d, e, the comparison of theoretical and experimental data in
                                                                             ð1Þ
    r1 ¼ λn1 Fk10 exp E 1
                             exp   αF ðU1 U10 Þ
                                                  C1m ¼ 8:97 ´ 104                     both seawater and pure water illustrate the accuracy and
                      RT                RT                                             feasibility of the thermodynamic–kinetic model.
where λ is rate ratio of the half reactions; n1 is the number of hole                     The variation trends of H2O2 generation over time are
transfer (n1 = 4); F is the Faraday constant (F = 96485 C mol−1);                      investigated in Fig. 4f. Specifically, no H2O2 is detected under
k10 is the rate constant of OER; E1 is the activation energy of                        the darkness in either pure water or seawater, but the yield of
OER; R is the gas constant (R = 8.314 J mol−1 K−1); T is the                           H2O2 is linear with time at the early stage of irradiation,
reaction temperature (T = 298.15 K); α is a constant (α = 0.5); U1                     suggesting that photocatalytic ORR is related to light. However,
is a potential (vs. RHE) in the Tafel region; U10 is the potential of                  after 72 h reaction, the yield of H2O2 almost remains unchanged,
water oxidation (U10 = 1.23 V); C1m is the concentration of                            and even shows a slight downward trend, which may be
hydroxyl of water (m = 1). Then, for ORR in seawater, the                              attributed to the decrease of catalytic activity and/or the
reaction rate r2 can be obtained.                                                      decomposition of H2O2. In addition, the aggregation experiments
                                                                                       in Fig. S19 suggest that the catalyst aggregate does not appear
                         r2
O2 þ 2H2 O þ 2e ! H2 O2 þ 2OH                                                        when the salt concentration in the range of seawater.
                                                                                      The re-usability of PM-CDs-30 has been studied via photo-
                                 E2         αF ðU2  U20 Þ n
               r2 ¼ n2 Fk20 exp        exp                   C2                        catalytic experiments using recovered samples. As shown in
                                  RT              RT                                   Fig. 4g, after five cycles of photocatalytic reaction, the H2O2
                                         ω 0:5                                       formation rate of PM-CDs-30 still remains at 1725 μmol g−1 h−1,
                  ¼ 2:31 ´ 1011 ´ PO2 ´           ðω > 0Þ                              indicating that PM-CDs-30 exhibits high stability in seawater.
                                         1600
                                                             ð2Þ                       The excellent stability of PM-CDs-30 was further corroborated by
                                                                                       electron microscopy and XPS analysis. The SEM image (Fig. S20a)
where n2 is the number of hole transfer (n2 = 2); k20 is the rate                      shows no obvious changes in morphology, while the TEM image
constant of ORR; E2 is the activation energy of ORR; U2 is a                           (Fig. 20b) indicates a slight aggregation of CDs after long times
potential (vs. RHE) in the Tafel region; U20 is the potential of                       photocatalytic reaction. Moreover, there are many signal peaks of
water oxidation (U20 = 0.68 V); C2n is the partial pressure of                         elements in the full XPS spectrum after the catalytic reaction
oxygen (n = 1); ω is the rotation speed in the reaction system                         (Fig. S21a), such as Na and Cl, compared with those before the
(r.p.m.); PO2 is the partial pressure of oxygen in the reaction                        catalytic reaction, which are mainly caused by the adsorbed ions
system. In particular, the rotational speed here is greater than                       and impurities from seawater. While, the C 1s and O 1s spectra of
zero. Next, the reaction rate (r3) of oxygen diffusion in ORR in                       PM-CDs-30 show no obvious change after continuous photo-
seawater is not negligible.                                                            catalytic reaction (Fig. S21b, S21c).
                           ω 0:8                           ω 0:8                      The AQY was measured to evaluate the light utilization
r3 ¼ Jmax ð60 r:p:m:Þ ´              ´ PO2 ¼ 5:24 ´ 103 ´              ´ PO2 ðω > 0Þ   efficiency of PM-CDs-30 under the illumination of monochro-
                              60                             60
                                                                                ð3Þ    matic light. As the action spectrum shown in Fig. 4h, the AQY
                                                                                       accords closely with the absorption spectrum of the photocatalyst,
   By comparing all the reaction rates of the photocatalytic                           demonstrating that the band gap is excited to generate H2O2.
reaction under certain conditions, the rate-limiting step is                           According to the equation in the experimental section, the AQY
obtained, that is, the reaction rate under this condition.                             of PM-CDs-30 can still reach 0.54% under the excitation light of
Depending on the above calculation, the thermodynamic–kinetic                          630 nm in seawater. Therefore, the catalyst can convert chemical
 model of PM-CDs-30 freely varying with rotational speed and                           energy well because the visible and near-infrared light account for
oxygen partial pressure in seawater is displayed in Fig. S16a,                         the most of the solar energy.
indicating that the reaction rate increases with the increase of
rotational speed, which is a gradual process. Meanwhile, within
the calculation range, the reaction rate increases continuously                        Proposed photocatalytic mechanism. The electron chemical
with the improvement of oxygen partial pressure. It should be                          impedance spectroscopy (EIS) technique was used to probe the
noted that, the rate of OER (r1) has no effect on the production                       charge-transfer properties of the catalysts (PM-CDs-x, Fig. S22a).
rate of H2O2 production. This is mainly attributed to the fact that                    The smaller semicircle for PM-CDs-30 indicates a faster electron
ORR and OER are parallel reactions in this photocatalytic                              transfer than others, suggesting that PM-CDs-30 possess the best
process, where the ORR process mainly relies on the dissolved                          migration property30. Time-dependent photo-response curves of
oxygen in seawater from air.                                                           these samples are tested with intermittent visible light (λ ≥
   Figure 4d shows the experimental and theoretical data on the                        420 nm) at open-circuit voltage in seawater (Fig. S22b). It indi-
relationship between catalytic activity and rotational speed in                        cates that PM-CDs-30 exhibits the best photo-response property,
both seawater and pure water. Similar to the calculation, the yield                    while PM-CDs-0 shows the worst compared to others, strongly
rate of H2O2 increased with the increase of the rotational speed.                      suggesting that PM-CDs-30 has superior charge generation
The same situation is also applicable to Fig. 4e, the theoretical and                  property compared to pristine polymer PM-CDs-031. A series of
experimental comparison of oxygen partial pressure and H2O2                            in situ TPV tests on PM-CDs-x (from PM-CDs-0 to PM-CDs-30)
yield, where the relationship between catalytic activity and oxygen                    were carried out and shown in Fig. S22c. A closer observation
partial pressure is consistent with the calculated results. It should                  reveals that the curve of the catalyst with a small amount of CDs
be pointed out that experimental data obtained is lower than the                       content (PM-CDs-10) decays slower than the curve of catalyst
calculated results because that the decomposition of H2O2 is not                       without CDs (PM-CDs-0). As the increase of CDs content, the
taken into account in the theoretical calculation. For comparison,                     curves decay to a negative value region, which expands as the CDs
the photocatalytic rates of catalyst in pure water were also                           content increases. All these results suggest that CDs promote
calculated and verified. As shown in Fig. S16b, the variation trend                     charge transfer at the interface of polymer and CDs, and effec-
of PM-CDs-30 reaction rate in water with rotation speed and                            tively improve the light absorption, charge carrier separation and
oxygen partial pressure is similar to that in seawater, but the                        conductivity of the catalyst. Also, CDs increase the electron sink

6                 NATURE COMMUNICATIONS | (2021)12:483 | https://doi.org/10.1038/s41467-020-20823-8 | www.nature.com/naturecommunications
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-20823-8                                                                        ARTICLE

Fig. 5 Photocatalytic mechanism of PM-CDs-30. TPV curves of a PM-CDs-0, b PM-CDs-30 before and after adding NaCl. c The photocatalytic H2O2
production rates upon the addition of different sacrificial agents. d EPR diagrams of PM-CDs-30 under darkness and light. e The production rate of H2O2 by
photocatalytic reaction in different gas environments. f TPV curves of PM-CDs-30 under different conditions. g H2O2 production rate varies with ion
concentration in different salt solutions. *P < 0.10; **P < 0.05; ***P < 0.01. The vertical error bars indicate the maximum and minimum values obtained; the
dot represents the average value.

barrier and electron trap ability of photoelectron and further                  the electron sink effect to “trapped” electron transfer, thus
inhibit the recombination of electron–hole pairs in PM-CDs-x.                   attracting more electrons from the detection system. In the
   Systematic TPV experiments of the catalytic systems with (PM-                presence of CDs, the TPV curves of PM-CDs-30 mixed with
CDs-30) and without CDs (PM-CDs-0) were carried out to                          other salt ions (Fig. S25), such as MgCl2 and CaCl2, also show
explore the mechanism of enhanced catalytic performance of salt                 similar phenomenon to NaCl. In addition, the difference
ions and CDs. The catalyst (50 mg) and NaCl (0.003 mol) were                    significance test of statistics method was adopted in the final
dissolved and lyophilized. Then the uniformly mixed sample was                  analysis, indicating that the cations can promote the production
excited under the same conditions as the original catalyst PM-                  of hydrogen peroxide in seawater.
CDs-0 and PM-CDs-30. As shown in Fig. 5a, the catalyst mixed                       To understand the mechanism of photocatalytic H2O2
with NaCl produced more charges than the original catalyst, and                 production over the PM-CDs-30 in seawater, by using AgNO3,
the negative region generated by charges recombination in                       EDTA-2Na, tert-butyl alcohol (TBA), and benzoquinone (BQ) as
attenuation was expanded. However, as a comparison, PM-CDs-0                    electron (e−), hole (h+), hydroxyl radical (·OH), and superoxide
(Fig. 5b) not only does not promote the generation of                           radical (·O2−) scavengers, the active species trapping experiments
photoexcited charges, but also does not appear in the negative                  were performed4,31. As shown in Fig. 5c, when AgNO3 is added
region during the decay time, indicating that the generation of                 to the photocatalytic reaction system, the yield of H2O2 decreases
negative region is mainly caused by the action of CDs in the                    sharply, indicating that the photogenerated electrons play a vital
polymer. The most likely reason for the negative region enlarging               role in the photocatalytic ORR. Separately, when EDTA-2Na is
in the decay process is that the ionization of CDs by Na+ increase              added to the system, the equilibrium concentration of H2O2 is

NATURE COMMUNICATIONS | (2021)12:483 | https://doi.org/10.1038/s41467-020-20823-8 | www.nature.com/naturecommunications                                   7
ARTICLE                                                                 NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-20823-8

increased, which is attributed to that the trapping of holes
promotes the use of electrons. Noting that little H2O2 is detected
when added BQ to the system, implying that the ·O2− is an
indispensable part during the process of photosynthetic H2O2. In
contrast, the addition of TBA shows almost no influence on H2O2
production, this is mean that ·OH did not participate in the
photocatalytic reaction. According to the previous reports32,33, we
conclude that PM-CDs-30 reduces oxygen to produce H2O2
through a two-electron reaction (O2 → ·O2− → H2O2), while
oxidizes water to releases O2 through a four-electron reaction
(H2O → O2).
   EPR detection was employed to record hydroxyl radical (·OH)
and superoxide radical (·O2−) generation during the photo-
catalytic reaction process in seawater. As shown in Figure S26,
there is no signal of DMPO-·OH observed under darkness and
illumination, confirming that no ·OH are produced in photo-
catalytic reaction. That is, the half reaction in which the holes
participate, water oxidation reaction, does not produce H2O2.
Notably, six DMPO-·O2− signals are observed after light
illumination in Figure 5d, suggesting that ·O2− is the main free
radical in the photocatalytic reaction. Based on the above tests
and experiments, it can be seen that the generated H2O2 is
obtained by the ORR involving electrons, while the water
oxidation process involving holes only releases O2 without any
H2O2 generation.
   A typical atmosphere course for H2O2 production over PM-                   Fig. 6 Schematic mechanism. Schematic mechanism of H2O2 production
CDs-30 under visible light irradiation is displayed in Fig. 5e. In            over PM-CDs-30 through photocatalytic reaction in seawater.
N2-saturated seawater obtained by bubbling N2 for 20 min, a
sharp decreased in photocatalytic H2O2 generation is observed.                seawater, indicating a four-electron reaction process for water
Different from the N2 environment, the production rate of H2O2                oxygen37. Thus, the photocatalytic reaction of PM-CDs-30 in
increased in seawater saturated with O2, showing that O2 was                  seawater is a two-channel parallel cascade reaction, that is, a
involved in the catalytic reaction. This result largely indicates that        parallel combination of two-channel oxygen reduction and four-
the photocatalyst PM-CDs-30 catalyzed ORR to synthesise H2O2                  electron water oxidation reactions.
and water oxidation to produce O2.
   A series of in situ TPV tests on PM-CDs-30 were performed
under different conditions (Fig. 5f). Obviously, the photovoltage             Discussion
intensity of PM-CDs-30 in O2-saturated acetonitrile decreases                 Based on above results and analysis, the possible mechanism for
sharply compared to N2-saturated acetonitrile, but increases                  photocatalysis on PM-CDs-30 is illustrated in Fig. 6. Under the
greatly when slight water is added. Here, the added O2 interacts              illumination of visible light (λ ≥ 420 nm), PM-CDs-30 can easily
with electrons in the ORR, reducing the intensity of the detected             absorb visible light to generate electron–hole pairs. The photo-
signal, while the added water, as a hole sacrificial agent, increasing         generated electrons are quickly transferred to the CDs, while
the amounts of electrons relatively. That is to say, in the                   photogenerated holes remain on the polymer. The electrons on
photocatalytic reaction, electrons are consumed by oxygen to                  the surface of CDs combine with oxygen in water to form H2O2,
produce hydrogen peroxide, while holes combined with water to                 and the holes in the polymer oxidize water to produce O2, which
release oxygen. Interesting, when water is substituted by seawater,           are performed simultaneously as two parallel reactions. As shown
the photovoltage shows the highest intensity, proving that the salt           in Fig. 6, due to the presence of metal cations in seawater, the
ions in seawater have a certain promotion effect on PM-CDs-30.                functional groups on the surface of CDs are ionized, which
   To compare the effects of the main components in seawater on               enhances the electron sink effect of CDs on electrons. Therefore,
catalytic activity, the synthetic efficiencies of H2O2 in three                CDs can attract and restrain more electrons, which prolongs the
separate ionic salts (NaCl, MgCl2, and CaCl2) were plotted in                 separation time of electron and hole, improves the catalytic
Fig. 5g. For this study, we controlled the concentration of salt ions         activity of the photocatalyst for photocatalytic oxygen reduction
around their respective concentrations in seawater, according to              to produce H2O2 in real seawater.
the international standard of seawater composition34. It should be               In this work, we demonstrate a metal-free composite catalyst
noted that, in the range of seawater concentration, the addition of           PM-CDs-30 with high photocatalytic activity for H2O2 produc-
NaCl, MgCl2, and CaCl2 improve the H2O2 yield of the                          tion in seawater. The superior performance of the photocatalyst
photocatalyst.                                                                originates from the addition of CDs in the catalyst, which
   The electron transfer numbers of PM-CDs-30 during the                      increases the time of electron–hole separation. Importantly, the
photocatalytic half reaction were evaluated by the rotating disc              ions in seawater ionize the functional groups on the surface of
electrode (RDE) and rotating-ring disc electrode (RRDE) in                    CDs, which amplifies the electron sink effect of CDs, making the
seawater. As depicted in Fig. S27, the electron transfer number is            photocatalytic activity in seawater better than in the pure water.
calculated by H2O2 detected in O2-saturated seawater, which is                PM-CDs-30 photogenerated H2O2 at a rate of 1776 μmol g−1 h−1
calculated to be 1.82 according to the equation mentioned in the              in real seawater, which is 4.8 times than that of the pure polymer
supporting information, suggesting a predominant two-electron                 (PM-CDs-0). The SCC efficiency of PM-CDs-30 can reach 0.21%,
pathway for ORR35,36. However, the ring current (red line) in                 which is an important breakthrough in the field of seawater
Fig. S28 does not change, while the disk current saliently                    photocatalysis synthesis of hydrogen peroxide. Moreover, theo-
improved when the light was applied to the N2-saturated                       retical calculations are consistent with the experiments, which

8                NATURE COMMUNICATIONS | (2021)12:483 | https://doi.org/10.1038/s41467-020-20823-8 | www.nature.com/naturecommunications
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-20823-8                                                                                             ARTICLE

give the optimum reaction conditions and the maximum reaction                             6 h under light irradiation, and then gas chromatography was used to detect the
rate under these conditions. Our catalytic system not only pro-                           change of oxygen.
                                                                                              At the end of the experiment, the solution was centrifuged, filtered, and then the
vides an idea for photocatalytic H2O2 production in seawater, but                         yield of H2O2 was determined by 0.1 M KMnO4.
also provides a promising way for the design, selection and
optimization of photocatalysts.
                                                                                          Data availability
                                                                                          The experimental data that support the findings of this study are available from the
Methods                                                                                   corresponding author upon reasonable request. Source data are provided with this paper.
Materials. All chemicals were not extra purified before use. Na2CO3 (AR, 98%)
and HNO3 (AR, 98%) were purchased from Aladdin. Procyanidins (PC, AR, 98%)
was purchased from Yuanye. 4-Methoxybenzaldehyde (MB, AR, 98%) was pur-                   Received: 20 August 2020; Accepted: 22 December 2020;
chased from Sinopharm Group chemical Reagent Co., Ltd. The seawater was taken
from the Yellow Sea.

Synthesis of CDs. The CDs were fabricated by an electrochemical method20,25.
Briefly, two graphite rods were cleaned in deionized water by ultrasonic agitation
for 30 min for three times. Then, the cleaned graphite rods were fixed as cathode          References
and anode, respectively, in the ultrapure water. Next, 30 V was applied across the        1.    de L. e Freitas, F. L. et al. Tunable catalytic performance of palladium
electrodes using a direct current (DC) power supply. The electrolytic process was               nanoparticles for H2O2 direct synthesis via surface-bound ligands. ACS Catal.
running for the 20 days under strongly stirring. After that, the black solution was             10, 5202–5207 (2020).
filtered with a four-stage membrane separation and filtration system (DMJ60) to             2.    Sun, J., Yu, Y., Curtze, A. E., Liang, X. & Wu, Y. Dye-sensitized photocathodes
obtain CDs solution with the particle diameter of less than 50 nm. Finally, the                 for oxygen reduction: efficient H2O2 production and aprotic redox reactions.
filtered CDs solution was freeze-dried to obtain CDs powder.                                     Chem. Sci. 10, 5519–5527 (2019).
                                                                                          3.    Xu, J. et al. Cd3(C3N3S3)2 coordination polymer/graphene nanoarchitectures
                                                                                                for enhanced photocatalytic H2O2 production under visible light. Sci. Bull. 62,
Synthesis of PM-CDs-x catalysts. The organic polymer composite systems PM-                      610–618 (2017).
CDs-x were synthesized according to the literature by a combination of sol–gel,           4.    Hu, S. et al. Photocatalytic oxygen reduction to hydrogen peroxide over
crosslinking polymerization and freeze-drying technology33. First of all, 0, 10, 30,
                                                                                                copper doped graphitic carbon nitride hollow microsphere: the effect of Cu(I)-
60, and 100 mg of CDs powders were respectively dissolved in 10 mL water to
                                                                                                N active sites. Chem. Eng. J. 334, 410–418 (2018).
obtain the CDs solutions of different concentrations. Then, 0.5 g of PC and 0.1 g of
                                                                                          5.    Kim, S. et al. Selective charge transfer to dioxygen on KPF6-modified carbon
Na2CO3 were dissolved in 10 mL CDs solutions, and then 0.9 mL of MB was added.
                                                                                                nitride for photocatalytic synthesis of H2O2 under visible light. J. Catal. 357,
After that, the mixture was sealed and transferred to an oil bath at 80 °C for 48 h to
                                                                                                51–58 (2018).
form a gelatinous precursor. The precursor was washed with ethanol and water to
remove residual organic solvents. Finally, after freezing with liquid nitrogen            6.    Liu, Y., Chen, F., Wang, Q., Wang, J. & Wang, J. Direct unassisted hydrogen
freezing (−196 oC, 10 min) and freeze-drying in a vacuum chamber (−53 oC,                       peroxide generation from oxygen and water on plasmonic Ag-graphene-Cu
0.113 mbar), the crosslinked organic semiconductor photocatalyst PM-CDs-x were                  nanosandwitch. Appl. Catal. B Environ. 224, 940–950 (2018).
obtained (x is the weight of the added CDs), which were successively labeled as           7.    Shiraishi, Y. et al. Resorcinol–formaldehyde resins as metal-free
PM-CDs-0, PM-CDs-10, PM-CDs-30, PM-CDs-60, and PM-CDs-100.                                      semiconductor photocatalysts for solar-to-hydrogen peroxide energy
                                                                                                conversion. Nat. Mater. 18, 985–993 (2019).
                                                                                          8.    Gryszel, M. et al. General observation of photocatalytic oxygen reduction to
Experiment of photocatalysis. The photocatalytic activity of the photocatalysts                 hydrogen peroxide by organic semiconductor thin films and colloidal crystals.
was assessed by a multichannel photochemical reaction system (PCX-50B, Beijing                  ACS Appl. Mater. Interfaces 10, 13253–13257 (2018).
Perfectlight Co., Ltd, China) equipped with a visible light source (λ ≥ 420 nm),          9.    Fukuzumi, S., Lee, Y.-M. & Nam, W. Fuel production from seawater and fuel
without the addition of sacrificial agents and other cocatalysts. In a typical reaction,         cells using seawater. ChemSusChem 10, 4264–4276 (2017).
10 mg of the catalyst was dispersed in a quartz bottle (50 mL) containing 20 mL of        10.   Mase, K., Yoneda, M., Yamada, Y. & Fukuzumi, S. Seawater usable for
seawater and then it was exposed to visible light for 5 h without seal in air at room           production and consumption of hydrogen peroxide as a solar fuel. Nat.
temperature. The rotation speed was 180 r.p.m. The optimal activity and reaction
                                                                                                Commun. 7, 11470 (2016).
mechanism of the catalyst were explored by changing the light condition (with and
                                                                                          11.   Hirakawa, H. et al. Au nanoparticles supported on BiVO4: effective inorganic
without light), the illumination time (5, 12, 24, 48, 72, 84, and 96 h), reaction
                                                                                                photocatalysts for H2O2 production from water and O2 under visible light.
environment (water or seawater), the excitation wavelength (λ = 365, 420, 535, and
                                                                                                ACS Catal. 6, 4976–4982 (2016).
620 nm), and the sacrificial agent (AgNO3, EDTA-2Na, TBA, and BQ) (2 mM).
                                                                                          12.   Moon, G. et al. Eco-friendly photochemical production of H2O2 through O2
After each reaction, the catalyst was washed and dried for stability testing (5 h for
each cycle).                                                                                    reduction over carbon nitride frameworks incorporated with multiple
    Experiments on the influence of rotating speed on catalytic activity were carried            heteroelements. ACS Catal. 7, 2886–2895 (2017).
out at different rotating speeds, which were 40, 80, 120, 160, and 200 r.p.m.,            13.   Baran, T., Wojtyła, S., Minguzzi, A., Rondinini, S. & Vertova, A.
respectively. Here, 10 mg catalyst was dispersed in 20 mL seawater (or pure water),             Achieving efficient H2O2 production by a visible-light absorbing,
and the quartz bottle (50 mL) was not sealed. The catalytic reaction took places in             highly stable photosensitized TiO2. Appl. Catal. B Environ. 244, 303–312
air at room temperature under visible light irradiation (λ ≥ 420 nm) for 5 h. For the           (2019).
experiment of the relationship between oxygen partial pressure and catalytic              14.   Zhu, Z., Pan, H., Murugananthan, M., Gong, J. & Zhang, Y. Visible light-
activity, the catalytic system was sealed, and then oxygen (0, 4, 8, 12, and 16 mL)             driven photocatalytically active g-C3N4 material for enhanced generation of
was injected into it to control the oxygen partial pressure of the catalytic system.            H2O2. Appl. Catal. B Environ. 232, 19–25 (2018).
Then, the catalytic reaction was carried out at 180 r.p.m. under visible light            15.   Zheng, Y. et al. Black phosphorus and polymeric carbon nitride
irradiation (λ ≥ 420 nm) for 5 h.                                                               heterostructure for photoinduced molecular oxygen activation. Adv. Funct.
    For the reaction atmosphere, N2 (or O2) was used to bubble in seawater for 20               Mater. 28, 1705407 (2018).
min, and then the catalytic activity of the catalyst was tested under different           16.   Zhao, S. et al. Carbon nanotubes covalent combined with graphitic carbon
atmospheres under airtight condition. As for the effect of different types of salt on           nitride for photocatalytic hydrogen peroxide production under visible light.
the activity, different concentrations of NaCl solution, MgCl2 solution, and CaCl2              Appl. Catal. B Environ. 224, 725–732 (2018).
solution were employed (the concentrations were within the range of seawater              17.   Shiraishi, Y. et al. Highly selective production of hydrogen peroxide on
concentration). Here, difference significance test of statistics method was adopted              graphitic carbon nitride (g-C3N4) photocatalyst activated by visible light. ACS
in the final analysis. The experiments were analyzed using one-way analysis of                   Catal. 4, 774–780 (2014).
variance (ANOVA) followed by Tukey–Kramer post hoc test, and the analytical               18.   Yang, X. et al. Water-soluble conjugated molecule for solar-driven hydrogen
and experimental replicates were used three time at least.                                      evolution from salt water. Adv. Funct. Mater. 29, 1808156 (2019).
    In order to detect the change of hydrogen peroxide yield with temperature, 20         19.   Zhu, C. et al. Construction of CDs/CdS photocatalysts for stable and efficient
mL seawater containing 10 mg PM-CDs-30 was performed at different                               hydrogen production in water and seawater. Appl. Catal. B Environ. 242,
temperatures (20, 40, 60, and 80 °C) under visible light irradiation for 5 h. In                178–185 (2019).
addition, the temperature change of the reaction solution (20 mL seawater with 10         20.   Zhu, M. et al. Strong coupling effect at the interface of cobalt phosphate-
mg catalyst) during the photocatalytic reaction was monitored by an infrared                    carbon dots boost photocatalytic water splitting. J. Colloid Interface Sci. 530,
camera under room temperature and unsealed conditions.                                          256–263 (2018).
    To monitor the evolution of O2, 10 mg of the catalyst was dispersed in a quartz       21.   Liu, J. et al. Metal-free efficient photocatalyst for stable visible water splitting
bottle (50 mL) containing 20 mL of water and then 10 mM of AgCl was added. The                  via a two-electron pathway. Science 347, 970–974 (2015).
system was bubbled by N2 for 30 min. The sealed reaction system reacted 2, 4, and

NATURE COMMUNICATIONS | (2021)12:483 | https://doi.org/10.1038/s41467-020-20823-8 | www.nature.com/naturecommunications                                                          9
ARTICLE                                                                          NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-20823-8

22. Lu, Y. et al. Spatial heterojunction in nanostructured TiO2 and its cascade        Foundation of China (51821002), National Natural Science Foundation of China
    effect for efficient photocatalysis. Nano Lett. 20, 3122–3129 (2020).               (51725204, 21771132, 51972216, 52041202), Natural Science Foundation of Jiangsu
23. Liu, C. et al. All-solid-state Z-scheme system of NiO/CDs/BiVO4 for visible        Province (BK20190041), Key-Area Research and Development Program of GuangDong
    light-driven efficient overall water splitting. Chem. Eng. J. 358, 134–142          Province (2019B010933001), Collaborative Innovation Center of Suzhou Nano Science &
    (2019).                                                                            Technology, the Priority Academic Program Development of Jiangsu Higher Education
24. Melchionna, M. & Fornasiero, P. Updates on the roadmap for photocatalysis.         Institutions (PAPD), and the 111 Project. Finally, we would like to thank Prof. Haiping
    ACS Catal. 10, 5493–5501 (2020).                                                   Lin for his contribution to the calculation part of this work.
25. Sun, Y. et al. A Pt-Co3O4-CD electrocatalyst with enhanced electrocatalytic
    performance and resistance to CO poisoning achieved by carbon dots and
    Co3O4 for direct methanol fuel cells. Nanoscale 9, 5467–5474 (2017).
                                                                                       Author contributions
                                                                                       Q.W., J.C., and Z.K. designed and planned the experiments. Q.W., X.W., and Y.L. pre-
26. Lin, X. et al. Evolution of oxygen vacancies in MnOx-CeO2 mixed oxides for
                                                                                       pared the samples. Y.Z, H.W., H.H., and L.F. carried out the photochemical experiment.
    soot oxidation. Appl. Catal. B Environ. 223, 91–102 (2018).
                                                                                       Q.W., J.C., H.H, Y.L., M.S., and Z.K analyzed the data. All the authors participated in the
27. Li, Y., Ho, W., Lv, K., Zhu, B. & Lee, S. C. Carbon vacancy-induced
                                                                                       interpretation and discussion of the results. Q.W., Y.L., M.S., and Z.K. wrote the article.
    enhancement of the visible light-driven photocatalytic oxidation of NO over
                                                                                       All co-authors contributed to the manuscript.
    g-C3N4 nanosheets. Appl. Surf. Sci. 430, 380–389 (2018).
28. Yang, Y. et al. A low-cost and stable Fe2O3/C-TiO2 system design for highly
    efficient photocatalytic H2 production from seawater. Catal. Commun. 143,           Competing interests
    106047 (2020).                                                                     There are no conflicts of interest to declare.
29. Liu, Y., Liao, Z., Ma, X. & Xiang, Z. Ultrastable and efficient visible-light-
    driven hydrogen production based on donor-acceptor copolymerized covalent
    organic polymer. ACS Appl. Mater. Interfaces 10, 30698–30705 (2018).               Additional information
30. He, Y. et al. Synthesis of Co(II)-Fe(III) hydroxide nanocones with mixed           Supplementary information is available for this paper at https://doi.org/10.1038/s41467-
    octahedral/tetrahedral coordination toward efficient electrocatalysis. Chem.        020-20823-8.
    Mater. 32, 4232–4240 (2020).
31. Hou, T. et al. Fe single-atom catalyst for visible-light-driven photofixation of    Correspondence and requests for materials should be addressed to Y.L., M.S. or Z.K.
    nitrogen sensitized by triphenylphosphine and sodium iodide. ACS Catal. 10,
    5502–5510 (2020).                                                                  Peer review information Nature Communications thanks Francesca Arcudi and other,
32. Luo, L., Wang, Z., Xiang, X., Yan, D. & Ye, J. Selective activation of benzyl      anonymous, reviewers for their contributions to the peer review of this work. Peer review
    alcohol coupled with photoelectrochemical water oxidation via a radical relay      reports are available.
    strategy. ACS Catal. 10, 4906–4913 (2020).
33. Wu, Q. et al. A function-switchable metal-free photocatalyst for the efficient      Reprints and permission information is available at http://www.nature.com/reprints
    and selective production of hydrogen and hydrogen peroxide. J. Mater. Chem.
    A 8, 11773–11780 (2020).                                                           Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
34. ASTM International. Practice for the preparation of substitute ocean water.        published maps and institutional affiliations.
    ASTM D1141-D1198 (2013).
35. Zhao, S. & Zhao, X. Polyoxometalates-derived metal oxides incorporated into
    graphitic carbon nitride framework for photocatalytic hydrogen peroxide                              Open Access This article is licensed under a Creative Commons
    production under visible light. J. Catal. 366, 98–106 (2018).                                        Attribution 4.0 International License, which permits use, sharing,
36. Zhao, S., Zhao, X., Zhang, H., Li, J. & Zhu, Y. Covalent combination of            adaptation, distribution and reproduction in any medium or format, as long as you give
    polyoxometalate and graphitic carbon nitride for light-driven hydrogen             appropriate credit to the original author(s) and the source, provide a link to the Creative
    peroxide production. Nano Energy 35, 405–414 (2017).                               Commons license, and indicate if changes were made. The images or other third party
37. Wu, Q. et al. Metal-free catalyst with large carbon defects for efficient direct    material in this article are included in the article’s Creative Commons license, unless
    overall water splitting in air at room pressure. ACS Appl. Mater. Interfaces 12,   indicated otherwise in a credit line to the material. If material is not included in the
    30280–30288 (2020).                                                                article’s Creative Commons license and your intended use is not permitted by statutory
                                                                                       regulation or exceeds the permitted use, you will need to obtain permission directly from
                                                                                       the copyright holder. To view a copy of this license, visit http://creativecommons.org/
Acknowledgements                                                                       licenses/by/4.0/.
This work is supported by National MCF Energy R&D Program (2018YFE0306105),
National Key Research and Development Project of China (2020YFA0406104 and
2017YFA0204800), Innovative Research Group Project of the National Natural Science     © The Author(s) 2021, corrected publication 2021

10                  NATURE COMMUNICATIONS | (2021)12:483 | https://doi.org/10.1038/s41467-020-20823-8 | www.nature.com/naturecommunications
You can also read